Pipeline Builder
Create and manage data pipelines with the visual drag-and-drop builder.
Overview
The Pipeline Builder uses React Flow to provide a canvas where you wire together nodes into data flows. Drag nodes from the palette, connect them, configure each node, and deploy.
Creating a Pipeline
- Navigate to Build > Pipelines in the sidebar
- Click Create Pipeline (or select a template)
- Drag nodes from the palette onto the canvas
- Connect nodes by dragging from output to input handles
- Configure each node by clicking on it
- Click Save Pipeline
Node Types
Source Nodes
| Type |
Description |
| AVEVA PI |
Connect to PI Server via PI Web API |
| OPC-UA |
OPC Unified Architecture protocol |
| Modbus |
Modbus TCP/RTU devices |
| MQTT |
MQTT broker subscription |
| Edge Agent |
Receive data from a FacilEdge agent |
| Airbyte |
300+ source connectors via PyAirbyte |
| Dataset |
Load a registered dataset as input |
Processor Nodes
| Type |
Description |
| Filter |
Filter data based on conditions |
| Transform |
Apply transformations (unit conversion, scaling) |
| Aggregate |
Roll up data (avg, min, max, sum) |
| Join |
Merge two streams by key or timestamp |
| Feature Engineer |
Compute derived features (rolling stats, lag, ratios) |
Sink Nodes
| Type |
Description |
| InfluxDB |
Write to InfluxDB time-series database |
| Kafka |
Publish to Kafka topics |
| S3 |
Store in S3-compatible object storage |
| Dataset |
Write output as a versioned dataset |
ML Nodes
| Type |
Description |
| ML Inference |
Run predictions using a registered MLflow model |
Pipeline Templates
Pre-built templates for common patterns:
| Template |
Description |
| OPC-UA to InfluxDB |
Basic OT data collection |
| Multi-source Join |
Merge PI + OPC-UA streams |
| ML Scoring Pipeline |
Source -> Feature Engineer -> ML Inference -> Sink |
| Quality Monitoring |
Source -> Aggregate -> Quality rules -> Alert |
Select a template from the Create Pipeline dialog to start with a pre-wired flow.
Test Mode
Before deploying, validate your pipeline with test mode:
- Click Test in the toolbar
- The pipeline runs with a limited data sample
- Each node shows pass/fail status and sample output
- Fix issues before deploying to production
Execution Logs
Pipeline execution logs stream in real-time via SSE:
- Open a running pipeline and click the Logs tab
- Logs stream live as the pipeline processes data
- Filter by log level (info, warn, error)
- Download full log history as a file
Pipeline Operations
| Action |
Description |
| Start |
Begin data collection and processing |
| Stop |
Pause pipeline execution |
| Edit |
Modify pipeline configuration |
| Delete |
Remove pipeline permanently |
| Duplicate |
Clone pipeline with all configuration |
| Export |
Download pipeline as JSON |
| Import |
Upload a pipeline JSON file |
Keyboard Shortcuts
| Shortcut |
Action |
Ctrl+Z / Cmd+Z |
Undo |
Ctrl+Shift+Z / Cmd+Shift+Z |
Redo |
Ctrl+S / Cmd+S |
Save pipeline |
Ctrl+A / Cmd+A |
Select all nodes |
Delete / Backspace |
Delete selected nodes |
Ctrl+D / Cmd+D |
Duplicate selected nodes |
Ctrl+C / Cmd+C |
Copy selected nodes |
Ctrl+V / Cmd+V |
Paste nodes |
Export / Import
Export
- Open a pipeline
- Click Export in the toolbar menu
- Pipeline is downloaded as a
.json file including all node configurations and connections
Import
- Click Import Pipeline from the pipeline list
- Select a
.json file
- Review the imported pipeline on the canvas
- Save to persist